Synthesis and sample characterization
X- Ray Powder Diffraction
Powder diffraction patterns are recorded on the crushed samples, to determine which phases are present (qualitative analysis) and their relative abundance, through Rietveld refinement. A Bruker-AXS D8 Advance diffractometer (Bruker, Karlsruhe, Germany) equipped with an Si(Li) solid state detector is used.
Intensity measurements are run at 40 kV and 30 mA and are taken using CuK
radiation ( = 1.54178 Å) in steps of 0.02° over a 2 range from 10° to 80°, with a counting time of 10 sec per step. The identification of the phases is done with the EVA software (Bruker), and ICDD database. Only Ca,pyroxene, olivine, Co-akermanite and a silica phase, likely trydimite, are found. The pyroxene is found in all but the Co10 sample (nominal composition Co2Si2O6), where olivine and tridymite are present instead; the tridymite is found in all runs, whereas Co-akermanite is present only from Co0 to Co4, i.e. in Ca-richer starting materials.
Olivine instead is found by X-ray diffraction between Co2 and Co10, in Co-richer samples.
The identification of the silica phase is difficult as only one peak at 2 = 23° is present. Both cristobalite and tridymite can be fit, and the assignment to tridymite is given mostly from reference to previous papers (Masse et al., 1999) and to the silica phase diagram. Some tests are done also using cristobalite as silica phase in the Rietveld refinements, obtaining poorer fitting.
The determination of the unit cell parameters and the phase quantification is performed by the Rietveld method with the GSAS-EXPGUI software package by multiphase refinement (Larson and Von Dreele, 2000). Independent variables for the Rietveld refinement were: scale factor, zero point, nine coefficients of the shifted Chebyschev function to fit the background, unit cell dimensions, and separate scale
Synthesis and sample characterization
factors for the different phases. A pseudo-Voigt profile function was used, refining two Gaussian coefficients (Gv and Gw) and one Lorentzian (Lv). The Rietveld analysis was done using structural data from Kimata et al. (1983)(Kimata, 1983) for Co-akermanite, from Morimoto et al. (1974) (Morimoto et al., 1974) for Co-olivine, from Ghose et al. (1987) (Ghose et al., 1987) for Ca,Co-pyroxene and from Konnert et al. (1978) (Konnert and E, 1978) for tridymite. A good fit of the observed vs calculated patterns is always obtained, with 2 values varying between 3.1 and 1.5.
Figure 2.2 shows a part (10-40 2) of the X-Ray diffraction pattern for Co0, Co5 Co10 at 1000°C for 12 hours and an example of peak identification. The unit cell parameters and phase abundance are listed in Table 2.3. To note for all phases the relevant peaks are sharp, indicating sample homogeneity and low strain.
Fig. 2.2: Part of a X-Ray diffraction pattern for Co0, Co5 Co10 at 1000°C for 12 hours. The phase
identification for the major peaks is reported.
Synthesis and sample characterization
Sample Co0 Co0 Co0 Co1 Co2 Co3 Co4 Co5 Co5 Co5 Co6 Co7 Co8 Co9 Co10 Co10
Synthesis condition
Co fraction
(a.p.f.u.) 1 1 1 1.1 1.2 1.3 1.4 1.5 1.5 1.5 1.6 1.7 1.8 1.9 2 2
Time (h) 12 48 24 12 12 12 12 12 48 96 12 12 12 12 12 48
Temper (°C) 1000 1000 1150 1000 1000 1000 1000 1000 1000 1100 1000 1000 1000 1000 1000 1000
Pigmentphasecomposition
CaCoSi2O6
wt% 50.6 81.9 98.1 61.1 59.1 58.5 55.7 54.3 60.4 59.1 45.8 33.5 22.2 11.0
a (Å) 9.791(1) 9.794(1) 9.798(1) 9.794(2) 9.794(1) 9.792(1) 9.788(1) 9.791(1) 9.796(2) 9.793(1) 9.783(2) 9.793(2) 9.790(3) 9.793(2) b (Å) 8.961(1) 8.957(1) 8.961(2) 8.962(1) 8.963(2) 8.965(2) 8.961(2) 8.963(1) 8.967(2) 8.961(3) 8.958(1) 8.966(3) 8.963(2) 8.965(3) c (Å) 5.241(1) 5.243(1) 5.246(1) 5.242(1) 5.242(1) 5.242(1) 5.239(2) 5.241(2) 5.243(1) 5.244(3) 5.240(1) 5.241(1) 5.241(1) 5.244(1)
(°) 105.53(2) 105.50(2) 105.51(2) 105.55(2) 105.58(2) 105.60(3) 105.60(2) 105.64(2) 105.70(4) 105.70(2) 105.65(4) 105.67(4) 105.69(1) 105.74(4) V (Å3) 443.03(1) 443.20(1) 443.78(3) 443.24(1) 443.28 443.18(1) 443.57(3) 443.94(3) 443.34(6) 443.02(4) 442.17(5) 443.07(7) 442.75(7) 442.83(7) Ca2CoSi2O7
wt% 34.9 13.9 26.8 19 9.6 6.7
a (Å) 7.833(2) 7.828(2) 7.834(2) 7.836(2) 7.833(2) 7.833(2) c (Å) 5.021(2) 5.022(3) 5.023(2) 5.023(3) 5.023(2) 5.018(2) V (Å3) 308.11(3) 307.73(5) 308.30(5) 308.30(3) 308.21(5) 307.86(3) Co2SiO4
wt% 11 19.3 22.6 31.2 32.7 34.5 37.6 47 57.3 66.4 70.7 72.3
a (Å) 4.782(2) 4.784(3) 4.781(3) 4.783(2) 4.783(3) 4.784(4) 4.781(3) 4.783(2) 4.783(3) 4.783(4) 4.782(4) 4.781(2)
b (Å) 10.311(4) 10.311(2) 10.312(4) 10.314(4) 10.320(2) 10.319(5) 10.308(4) 10.316(4) 10.314(5) 10.313(2) 10.302(5) 10.301(4)
c (Å) 6.008(3) 6.007(3) 6.005(3) 6.008(3) 6.012(3) 6.011(3) 6.005(2) 6.009(4) 6.008(3) 6.008(2) 6.003(4) 6.002(2)
V (Å3) 296.46(5) 296.41(4) 296.06(5) 296.33(7) 296.80(5) 296.74(5) 295.91(4) 296.48(7) 296.35(7) 296.34(3) 295.73(9) 295.59(8) SiO2
wt% 14.4 4.0 1.5 11.9 10.8 12.5 14.9 12.2 6.1 6.0 16.4 19.4 20.3 22.5 29.2 27.6
a (Å) 4.915(5) 4.959(5) 4.956(5) 4.923(5) 4.894(5) 4.942(5) 5.007(5) 4.993(5) 4.980(5) 4.983(5) 4.985(5) 4.988(7) 4.988(6) 4.989(7) 4.989(7) 4.997(7) c (Å) 7.279(6) 7.157(4) 7.118(3) 7.271(7) 7.352(6) 7.173(5) 6.977(6) 7.003(6) 6.988(6) 6.976(7) 6.992(5) 7.006(7) 6.992(5) 6.987(5) 6.983(7) 6.943(5) V (Å3) 175.85(4) 176.03(4) 174.85(5) 176.19(4) 176.11(3) 175.17(5) 174.89(4) 174.62(3) 173.31(5) 173.19(3) 173.77(1) 174.28(7) 173.98(7) 173.95(3) 173.81(5) 173.39(5)
Tab.2.3: Cell parameters and wt fraction for some synthesized samples. The error for single measurement is reported in brackets. If the wt% fraction is less than 5% the Rietveld refinement doesn’t reveal exactly the phase content.
Synthesis and sample characterization
In spite of the large changes in bulk composition, cell parameters of the synthesized phases change very little. This indicates, as confirmed by subsequent SEM-EDS results (following chapter), that the composition of the single phases doesn’t change, and what varies is their abundance.
More in detail, in the Ca,Co-pyroxene phase the a, b, c parameters don’t change significantly, but we found a small change in the angle (105.53° - 105.74°) and in volume (443.7 Å3 - 442.8 Å3), which can be hardly related to changes in the bulk composition. Such small changes can be interpreted anticipating the results on the high pressure pyroxenes, whose composition is the same as the bulk starting materials (following chapter, Part III). From the plot in figure 2.3 it appears that the changes in and volume for the low pressure series are little more than an experimental and compositional variation around a Co content of about 0.1 - 0.2 a.p.f.u in M2. These small quantities of Co in M2 indicate that some solid solution of Co in the M2 site is possible also at room pressure, as will be confirmed by EDS analysis.
Fig. 2.3: angle (°) vs volume value (Å3) for the samples synthesized ad room pressure and high pressure. The variation of the parameters for the room pressure experiment correspond
to a composition at about 0.1 – 0.2 of Co in M2 (a.p.f.u.).
Synthesis and sample characterization
As regards Co-akermanite cell parameters are the same within error in all runs, indicating an homogeneous composition. Co-olivine cell volume is slightly different (0.3% higher) in Ca-bearing samples respect to the sample Co10 were Ca is not present in the starting materials. In Ca-bearing samples, where olivine coexists with pyroxene, the cell volumes of Co-olivine are instead very similar. The likely explanation is the presence of some Ca in solid solution in the olivine structure, which increases the cell volume by its higher than Co cation radius. Again this will be confirmed by SEM-EDS analyses.
Despite the problem of the poor definition of the silica phase the Rietveld refinement is successful in determining the abundance.
Figure 2.3 shows the trend of the weight percentages of the various compounds as a function of Co content in the bulk composition. We observe the major presence of Co-akermanite in Ca-richer samples, that gradually disappeared while cobalt increased, leading to the formation of Co-olivine. We note that Ca,Co-pyroxene is present, during almost the entire series (except on Co10 where there’s no Ca), but its abundance decreases drastically from Co5 onwards. The tridymite content increases in the Co-rich samples. It is also observed that in CaCoSi2O6, prolonging heating and increasing temperature the pyroxene phase is formed at expenses of Co-akermanite, that disappeares in longer annealing (figure 2.4).
Synthesis and sample characterization
Fig. 2.4: wt fraction (%) of the phases vs Co content (a.p.f.u.) in M2. The sample Co0 is annealed at different temperature and for different time: in this second step the
pyroxene phase is formed at expenses of Co-akermanite, that disappeares in longer annealing.
The assemblage obtained in the first series of runs and Ca-richer samples, does not represent equilibrium, as it can be shown by simple thermodynamic considerations. In a system with a T and P fixed the Gibbs’ phase rules establishes that the degree of freedom of a system (f) is:
f = c – p +m + 0 (P,T)
where c is the number of the components, here 3 (CaO, CoO, SiO2), p the number of the phases obtained, and m the number of external parameters that have an effect on the state of the system. In our case m, at a pressure and temperature constant equals to 0. Obviously the phase freedom can be at minimum 0, in an invariant system.
In Ca-richer samples (see table 2.3) we have found four phases, - Ca,Co,pyroxenes, Co-olivine (found in Raman spectra), Co-akerminite and tridymite -, so that the freedom is -1, which indicates that at least one of the obtained phases is metastable.
Synthesis and sample characterization
In a ternary diagram CoO, CaO, SiO2 , based on Mukhopadhyay et Jacob (1996) phase diagram, our bulk composition should lie in the ternary field included between Co2SiO4, SiO2 and a phase rich in Ca and Co, like pyroxene or Co-akermanite. The suggestion is that one either Co-akermanite and CaCO-pyroxene are not stable at the experimental condition, and evidence is that the metastable phase is Co-akermanite: while increasing annealing time at constant temperature the CaCo-pyroxene increases gradually at the expense of the Co-akermanite. The same happens if we increase temperature (table 2.3).
The single pyroxene phase, which is the final product of longer annealing and higher temperature is a stable phase at least above 1000°C for the CaCoSi2O6 starting materials. Co-akermanite can be a residual of a stability field that is present a lower temperature during the slow cooling in the furnace or simply a metastable phase formed during the first step of the synthesis.
To note, Co-akermanite is found in the synthesis of CaCoSi2o6 pyroxene also by Masse et al. (1999), who did not clarify the point of metastability.
Fig. 2.11: ternary diagram of the studied samples. The composition of the studied phases is shown, as well as that of the starting materials (small dots). Note that the starting materials
fall into a ternary field, determined by Co2SiO4-CaCoSi2O6-SiO2, or Co2SiO4-Ca2CoSi2O7 -SiO2. In the former case stoichiometric CaCoSi2O6, would give a pyroxene only assemblage.
Synthesis and sample characterization
Scanning Electron Microscopy and EDS microanalysis
The size and the shape of the different phases are observed through SEM backscattered images, and their composition is determined with microprobe EDS analysis. SEM-EDS analyses are performed with a Jeol 6400 Scanning Electron Microscope equipped with an Oxford EDS (Energy Dispersive System). Operating conditions are 20 kV and 1.2 mA current, and analyses are taken using ~1m beam diameter and 75 sec counting time. Cobalt is used as a standard. An average of about 15 spots per sample is obtained. The images are taken under low vacuum conditions. The samples are glued in epoxy resin and polished, to make them flat and covered with a high-conductance thin film (graphite film) to avoid charging effects. Analyses of the room pressure runs are done only on samples with bulk composition Co0, Co5 and Co10.
Figures from 2.4 to 2.11 show backscattered electron (BSE) images. The different grey tone is related to the average atomic number (Z) of the specimen: the darker corresponds to the lower number of backscattered electron, i.e. to the lower mean atomic number; the mean atomic numbers Z for the phases found in our syntheses are: Z = 14.5 for Co-olivine, Z = 10.6 for SiO2, Z = 12.7 for CaCoSi2O6 and Z=12.9 for Co-ackermanite. Co-ackermanite and pyroxene are hardly distinguishable in backscattered electron images, being their average atomic number similar.
Only in few cases they reached 5-10 µm. Large silica nuclei are present, with reaction rims of clinopyroxene, olivine and Co-akermanite.
Synthesis and sample characterization
In the first series of runs a common feature is that they are microcrystalline aggregates and mostly anhedral: crystals sized of the order of few microns; only in few cases they reached 5-10 µm. Large silica nuclei are present, with reaction rims of clinopyroxene, olivine and Co-akermanite
This situation is texturally indicative of a non-equilibrium, and it is mostly found in Co0. Further evidence of the non-equilibrium are rounded terminations of the crystals, or reaction rims. In figure 2.4 a reaction rim showing silica at the core, surrounded by Co pyroxene, mantled by Co-akermanite, is shown.
Fig. 2.4: BS image of the sample a)Co0, 1000°C*12h
b)Co5, 1000*12h c) Co10, 1000°C*12h
Synthesis and sample characterization
The above non-equilibrium textures are indicative that at the temperature of 1000°C for 12 hours the reaction is not completed and thermodynamic equilibrium is not reached. Increasing heating time and temperature larger crystals and more homogeneous samples with similar crystal size appear, like it is shown in figure 2.5:
the crystals are still tiny, sized at maximum 10 microns, but become euhedral and reaction rims disappear, as well as large tridymite crystals.
Fig. 2.6: BS image of the sample Co0: a)1100°C*96h and b)1150*24h Fig. 2.5 : BSE image of a small crytal in a no-equilibrium situation: a silica nucleus is still reacting, showing the formation of a pyroxene and Co-akermanite phases.
Synthesis and sample characterization
In fact we can see that the silica nuclei gradually reduced when temperature or time of the synthesis increased, although an excess silica was always present, related to an excess of SiO2 in the starting material.
Fig. 2.7: BS image of the sample Co0: a)1000°C*12h and b)1000*48h
As previously said, in the synthesis carried out at 1200°C, the samples are partially or totally molten after 3 and 24 hours annealing. The melting of samples caused a gradual leaching of cobalt that dissolved into the glass, changing the coordination in tetrahedral and giving a deep blue-black colour (figure 2.7).
Chemical analyses are performed with EDS only for Co0, Co5 and Co10 samples (table 2.4). The SEM-EDS analysis confirms the phases found by XRD.
A major analytical problem is the point analysis on overlapping crystals, due to their dimension close to the beam analyzed volume (few m3). Analysis whose stoichiometry cannot be classified in one of the reported phases are therefore discarded.
CaCo-pyroxene shows a different composition between samples Co0 and Co5. In fact in Co5 samples a Co-richer pyroxene (Ca0.9Co1.1Si2O6), coexisting with Co-olivine and tridymite in agreement of the suggestion of the powder XRD analysis of a slightly Co-richer solid solution at room pressure. As regards Co-olivine a small
Synthesis and sample characterization
Ca content (0.004 Ca atoms p.f.u.) is found in the analysis of the Co5 samples confirming the XRD diffraction results.
Table 2.4 shows chemical analysis made on the samples Co0, Co5 and Co10, at different temperature and time. Estimated error is on the last digit.
Sample Co0 Co0 Co0 Co0 Co0 Co0 Co5 Co5 Co5 Co5 Co5 Co10 Co10
Co nom. 1 1 1 1 1 1 1.5 1.5 1.5 1.5 1.5 2 2
Time (h) 12 48 120 3 96 24 12 48 3 24 96 12 48
Temp (°C) 1000 1000 1000 1100 1100 1150 1000 1000 1100 1100 1100 1000 1000
n° analysis 2 3 2 5 6 4 6 6 5 5 5
Cation on the basis of 6 O atoms per formula units (CaCoSi2O6)
Ca 1.04 0.96 0.98 0.95 0.93 1.03 0.88 0.85 0.8 0.9 0.83 Co 1.02 1.02 1.06 0.94 1.07 1.02 1.12 1.15 1.14 1.1 1.15 Si 1.97 2.04 1.96 1.99 1.98 1.98 1.99 1.98 2 1.99 1.98 Total 4.03 4.02 4.00 3.88 3.98 4.03 3.99 3.98 3.94 3.99 3.96
Cation on the basis of 7 O atoms per formula units (Ca2CoSi2O7) n° analysis
n.a.
1
Ca 1.91
Co 1.03
Si 2.01
Total 4.95
Cation on the basis of 4 O atoms per formula units (Co2SiO4) n° analysis
n.a. n.a.
3 2 2 1 5 3 4
Co 2 1.98 1.95 1.96 2.04 2 1.99
Ca n.d 0.01 0.02 0.01 0.01
Si 1.99 1.98 1.99 2.01 2 1.99 1.99
Total 3.99 3.97 3.96 3.98 4.05 3.99 3.98
Synthesis and sample characterization
Raman Spectroscopy
Raman spectroscopy is performed on the samples synthesized at room pressure to determine the phases present. By this respect it is complementary to X-ray phase determination. The information specific for Raman spectroscopy, i.e. an observation on roto-vibrational modes of molecules to identify phases but also to analyze the order/disorder into the crystal structure will be fully exploited in the analysis of the high pressure samples.
Raman spectroscopy is performed on powdered samples using a Jobin-Yvon Horiba LabRam microRaman apparatus, with spatial resolution of ~2 μm and a spectral resolution of ~2cm-1. The 632.8 nm line of an He-Ne laser is used as the excitation.
A filter wheel is used to reduce the laser power from 20 mW to 1 mW or less on the sample. The frequency calibration is performed against the Raman peak of silicon.
On each sample, an average of five Raman spectra are taken. The peak positions are obtained from baseline-corrected spectra by least-squares analysis using the computer program Labspec (HORIBA). A Lorentzian profile is used in the peak profile analysis.
Figure 2.8 shows Raman spectra of the samples synthesized at 1000°C for 12 hours. The spectra reveal sample heterogeneity: we find different peaks ascribable to different phases in the same spectrum. This is due to the presence of crystals sized close or less the area excited by laser spot, which, taking into account the penetration of the laser by few microns, is at least 5-10 µm3: as the crystals dimension of the room pressure runs are generally smaller the information is that of a sum of several crystals. Figure 2.6 shows a sketch of the area included in Raman analysis.
Synthesis and sample characterization
It should be also noted that the intensity of the peaks from the different phases vary locally, and has no meaning of average compositional content, differently from the IR spectroscopy.
To analyze the obtained spectra single phase standards are taken, and used to separate the different phases (figure 2.8 b). The most intense peaks for each phase are reported in table 2.5.
Raman data confirm the presence of the phases found by XRD, but also show the presence of Co-olivine in the Co0 sample. The phases present in the synthesis from the Co0 starting material of the first series of runs are the Co-akermanite, Ca,Co-pyroxene, Co-olivine and tridymite. The observation of Co-olivine in the Ca0 sample completes the XRD and SEM data: Rietveld refinement doesn’t reveal the presence of Co-olivine being this phase present in a small quantities (<5% in weight).
In the first series we find in all samples except in Co10 the typical pyroxene pattern:
the three peaks at 300-400 cm-1 and the two intense peaks, one at 660 cm-1 and the other at 1010 cm-1, which will be redetailed in the high pressure paragraph.
The peak near 900 cm-1, ascribable to Co-akermanite gradually decreases its intensity, disappearing at the composition of Co5, while the double peak at 810 and 840 cm-1 ascribable to the Co-olivine is clearly visible in all the series.
Synthesis and sample characterization
Fig. 2.8: a) Raman spectra of samples synthesized at 1000°C for 12 hours, b) standard used for the identification of the phases
Sample Temperature (°C)
Time (h)
ν1 Ca,Co-px
ν2 Ca,Co-px
ν3 Ca,Co-px
ν4 Ca,Co-px
ν5 Co-olivine
ν6 Co-olivine
ν7 Co-akerm
ν8 Ca,Co-px
Co0 1000 12 306.1 344.9 375.4 659.3 811.4 840.3 899.7 1009.7 Co0 1000 48 308.2 342.1 375.0 661.0 812.6 840.1 901.3 1009.4 Co0 1000 120 310.2 343.1 376.8 665.5 810.3 840.5 897.6 1010.4 Co0 1100 3 312.7 340.5 377.8 664.7 812.5 840.1 895.1 1013.2 Co0 1100 96 311.7 345.6 380.2 665.2 814.3 838.3 891.0 1011.2 Co0 1150 24 311.0 345.3 378.2 662.9 810.2 839.0 893.2 1011.5 Co1 1000 12 305.0 344.6 376.7 662.0 809.9 838.9 901.1 1010.0 Co2 1000 12 305.7 342.0 374.5 661.2 811.7 838.3 900.6 1008.8 Co3 1000 12 306.0 337.6 371.2 658.9 810.5 840.7 895.5 1006.4 Co4 1000 12 306.4 337.9 372.3 659.5 808.0 839.9 894.6 1007.0 Co5 1000 12 311.1 341.9 376.0 661.5 810.2 838.0 1009.8 Co5 1000 48 313.4 343.3 378.9 662.6 811.5 838.4 1010.5 Co5 1100 3 311.5 345.0 380.4 664.8 812.7 839.5 1012.9 Co5 1100 24 311.8 344.8 379.9 665.1 812.8 839.1 1013.2 Co5 1100 96 313.8 348.8 379.9 664.9 812.7 839.1 1013.2 Co6 1000 12 311.1 342.0 375.9 661.6 811.1 838.3 1009.9 Co7 1000 12 309.5 340.7 379.1 660.7 810.2 836.5 1009.1 Co8 1000 12 308.2 336.0 380.2 660.0 809.2 835.9 1008.6
Co9 1000 12 810.4 837.0
Co10 1000 12 811.0 837.1
Co10 1000 48 810.5 836.8
Table 2.5 peak position of the fitted Raman spectra with some attribution of the peaks. Error is estimated ± 0.5
Synthesis and sample characterization
In all runs the peak position of the phases does not change significantly, varying only their intensity: this confirms the observation of almost constant composition of the phases by XRD and SEM-EDS.
In figure 2.9 a the Raman spectra of the second series of runs on the sample Co0 are reported. The presence of characteristic peaks of the pyroxene is evident (the triplet between 300-400 cm-1, 660 and 1013 cm-1), but at the same time are visible peaks ascribable to Co-akermanite (900 cm-1) and Co-olivine (811 and 837 cm-1).
Increasing the annealing time, (1000°C x 120h) Co-olivine disappears but peaks of Co-akermanite are still present. Raising the temperature to 1100 °C, a typical spectrum of the pyroxene is observed, but two small peaks at about 620 cm-1 and 1000 cm-1 ascribable to Co-akermanite are also found. In the synthesis carried out at T = 1100° C for 96 hours and T = 1150°C for 24 hours the Raman spectrum reveals the presence of Ca,Co-pyroxene only. At 1200°C the sample begins to melt.
The Raman spectrum reveals the coexistence of glass and pyroxene, with larger peaks, due to inherent structural disorder.
The samples Co5 and Co10 show Co-olivine and tridymite in all runs, together with clinopyroxene in Co5 (figure 2.9 b,c).
Synthesis and sample characterization
Fig.2.9 : Raman spectra of the samples Co0, Co5 and Co10 with different synthesis condition.
Synthesis and sample characterization
Colour coordinates
The colour of a pigment provides the basis for any interest in manufacturing, but also conveys important analytical information.
Colour is the sensation that we perceive visually from the refraction or reflection of light on the surface of objects. Colour is light — as it is strictly related to it — and depending on the type of light (illuminating or luminous stimulus) we see one colour or another. Light is highly variable and so too is colour, to a certain extent.
Pigments absorb a part of the radiations of the incident light and reflect another, which reaches the eye of the observers, making them experience the sensation of colour. For instance, the dark colour of a pigment is almost entirely due to the fact that incident radiation is absorbed almost completely in the visible spectrum.
The purpose of colour analysis is to measure chromatic characteristics of the pigments in an objective way. This is done according to CIE standards.
An organization called CIE (Commission International de l'Eclairage) determines standard values that are used worldwide to measure colour. The values use by CIE are called L*, a* and b* colour coordinates and the colour measurement method is called CIELAB.
Fig. 2.10: CIELAB coordinate system
Synthesis and sample characterization
L*a*b* colour parameters can be measured with a spectrophotometer using a standard lighting C, following the CIE-L*a*b* colorimetric method.
The field of blue colour is mainly governed by the parameter b*: the more negative the b* value is, the bluer the colour hue results. On the other hand, the coordinate L*
gives us the lightness of the pigment (the higher L*, means lighter hue); it is also an indirect measurement of the brightness or intensity of the pigment (the lower L*
means the brighter or more intense the colour). In the analysis of our syntheses colorimetric analysis enables to assess the changes in colour and relate the colour components the composition of the different phases. The final colour can also be calculated from colour components, and represented as done in table 3.1.
The products of all the room pressure syntheses are measured with a colorimeter (spectrophotometer) with di/8 geometry (diffuse component seen at 8° from the normal to the sample), sampling step by 20 nm, measured between 360 and 780nm.
The colorimetric calculation is done after spectral reflection measurement with the standard CIE 1964 observer and Daylight.
The colour parameters (L*a*b*) of our samples are shown in Table 3.1.
In our samples the colour comes only from the Co-bearing phases, i.e. not from tridymite. The colours of the pure phases are deep blue for Co-ackermanite, (Ca2CoSi2O7) and pink for pyroxene (CaCoSi2O6) and olivine (Co2SiO4). Colorimetric analysis shows that some difference is present between the two pink phases: pure olivine, found in the Co10 synthesis shows lower luminance, a* and b* components vs pure CaCoSi2o6 clinopyroxene found in the longest synthesis at 1150°c. Colour coordinates of single phase pyroxene were 25 < a* < 27 and -10 < b* < -8. For the single Co-olivine phase the a* parameters has a similar value, but b* was more
Synthesis and sample characterization
negative (b* about -20). Smaller value in b* parameters indicate a more blue-violet hue. In fact the sample Co10 made by olivine exhibited an achromatic violet colour (very low a* and b* values), and the resulting colour of a single olivine phase appeared more violet respect to the pyroxene.
By mixing the different phases, the resulting colour reflects both colour efficiency and phase abundance. The most apparent colour changes are those in the first series runs from the (Co0) CaCoSi2O6 bulk composition to the (Co10), Co2Si2O6 where a change from blue to pink is observed. This is due to the presence of the Co-akermanite in the samples Ca-richer (from Co0 to Co5), that produced blue colour hues (b* values around 36).
Fig. 2.11 a) shows the b*parameters vs Co-akermanite content (wt% franction) for the samples from Co0 to Co5 (1000°C for 12hours) b) a*vs b* colour parameters for the samples
Co0 annealing at different temperature and time
In figures 2.11 a the colorimetric parameters b* vs Co-akermanite content obtained by Rietveld refinement are reported. The Co-akermanite is never present as dominating phase, but the colour for Co0 is deep blue, very similar to the colour of pure ackermanite. This occurs as cobalt in tetrahedral coordination has a more